Lung cancer is the leading cause of cancer-related death in the U.S general population as well as in the Veteran population. NSCLC (Non-Small Cell Lung Cancer) is the most common type of lung cancer, accounting for approximately 85% of all lung cancer cases. NSCLC is predominantly in an advanced stage (stage IIIB/IV) at the initial presentation. Conventional treatment for advanced NSCLC is to use systemic chemotherapy that kills cancerous cells but also has toxic effects on normal cells. Targeted agents are oral therapies that attack cancer cell-specific attributes. Anti-EGFR targeted agents, such as small molecule TKIs erlotinib (Tarceva). In general, they offer less severe side effects than chemotherapy but only offer significant benefit to a small subset of the population. Targeted agents are costly, and the effectiveness of EGFR TKIs in patients with an unknown or wild-type mutation has not been well- established. To our knowledge, there are no studies that assess the evidence-based benefits and values of delivering genomic-based targeted therapy relative to conventional therapies for lung cancer patients in VA healthcare system.
The goal of our study was to assess the utilization and comparative value of genomic-based targeted therapy (clinical effectiveness and cost-effectiveness), and oncologists' views of delivering genomic-based targeted therapy for Veterans with advanced lung cancer in a VA practice setting. Our specific aims were:
1.To examine the utilization of Genomic-Based Erlotinib (tarceva)Therapy (GBET) in the management of Veterans diagnosed with NSCLC at VHA.
2.To examine the clinical effectiveness of delivering GBET in Veterans with lung cancer in terms of overall survival and adverse events associated side effects of the treatment at VHA.
3.To examine the joint value of clinical effectiveness and cost of delivering GBET in management of Veterans with lung cancer relative to usual care.
4.To identify facilitators and barriers to the delivery of genomic-based targeted therapy for patients with lung cancer at the VHA. We conducted semi-structured interviews with medical oncologists at the VA eliciting information in four domains: (1) knowledge, (2) attitude, (3) intent and (4) barriers and facilitators for practicing genomic-based targeted therapy.
The project design was a retrospective, observational cohort study using existing VA clinical databases to identify Veterans diagnosed with advanced non-small cell lung cancer in FY 2010-2013. Study subjects are tracked from the index dates of their enrollment in the study until they die or are censored on September 30, 2015. Eligible patients were followed for up to 24 months. We used data from the Veterans Administrations Corporate Data Warehouse (CDW), including VA CDW data, VA Medical datasets, Decision Support System (DSS), and VA Central Cancer Registry (VACCR).
The study included 22,933 Veterans who met the inclusion and exclusion criteria and enrolled in the VA healthcare system diagnosed with primary lung cancer and histologically or cytologically confirmed non-small cell lung cancer between October 1, 2009 and September 30, 2013. Patients were excluded if they were 1) less than 21 years of age 2) diagnosed with secondary lung cancer 3) receiving hospice care within 30 days of lung cancer diagnosis 4) diagnosed with other comorbid conditions preventing active treatment of lung cancer 5) diagnosed with lung cancer by autopsy report 6) deceased within 30 days since the first lung cancer diagnosis date, and 7) non-Veterans.
We examined patients' ascertainment of GBET (Exposure to Treatment), ascertainment of non Erlotinib therapy (surgery, radiation, chemotherapy), ascertainment of EGFR testing and mutation status, outcome variables such as overall survival, cost of inpatient and outpatient care, and medications, overall lifetime total all-cause treatment cost and incidental cost per life-month gained, in addition to patient characteristics such as age, race and distance to seek care and hospital characteristics including comprehensive cancer center status.
To fulfill aim 4, we conducted one-time, semi-structured interviews with 30 oncologists involved in the treatment and management of patients with lung cancer in the VA.
Findings from our quantitative analysis
Among 22,933 veterans diagnosed with lung cancer at the VA between October 1, 2009-September 30, 2013, 1559 patients were prescribed Tarceva. Tarceva users were more often diagnosed at higher stages (stage 3b and 4: 64.53% vs 41.01%) compared to non-Tarceva users and were more likely to die during the follow up period of at least 24 months (77.23% compared to 56.34%). Tarceva users were more likely to receive chemotherapy (81.01% vs 37.63%) or any treatment (91.85% vs 77.21%) compared to non-Tarceva users but they were less likely to undergo surgery (16.74% vs. 31.85%). Non-Tarceva users were more likely to be white (75.80% vs 70.94%), with a history of smoking (58.80% vs 53.62) compared to Tarceva users and were less often female (2.44% vs 4.17%). The two groups had roughly the same distribution of age, marital status and use of drug or alcohol use. The mean survival was 527 days in patients of the Tarceva group while 419 days in the non-Tarceva group. The mean cost per patient-year was $111,352 in the Tarceva group and $127,098 in the non-Tarceva group.
In a multivariate logistic analysis of the receipt of tarceva, the factors that were statistically significant associated with use of Tarceva were receipt of chemotherapy, female gender and cancer stage 3 or 4, with odds ratios of 7.02, 1.79 and 1.53 respectively. Hazard ratios of mortality during the study period were for 0.845 (0.790 - 0.904) receipt vs nonreceipt of tarceva, when adjusted for stage, demographic factors, smoking history and receipt of other treatments (radiation, chemotherapy and surgery). In this case, the strongest adjusting factor was cancer stage with a hazard ratio = 4.220 for stage 3 or 4 vs lower stage.
In a multivariate regression analysis of cost, it showed lower overall cost for those receiving tarceva, when adjusted for demographic factors, smoking history, cancer stage and other confounding factors. Cost-effectiveness ratios showed an overall savings ($11462.94) associated with Tarceva receipt per extra survival month. A net cost savings showed cross all of individual stage from 1 to 4 as well.
Findings from our qualitative interviews
Testing decisions were based on disease characteristics, patient characteristics and evidence. Stage was an important factor influencing testing decisions. The majority of providers interviewed only reported testing patients with stage IV disease. The frequency of testing varied significantly, ranging from 5%-100% of metastatic non-small-cell adenocarcinoma patients. All providers agreed that mutations were rare in the VA but all believe the testing to be justified. The most common barriers to testing that providers encountered were limited sample, long wait times, no notification of results and needing to send out the sample for testing.
The providers that were interviewed describe guideline-concordant care. All the oncologists interviewed had experience with targeted therapies, in particular, with erlotinib. Factors that influence their prescribing decisions include patient characteristics such as social support, proximity to the VA and performance status. Most providers endorsed no barriers to prescribing erlotinib. Most providers interviewed believe targeted therapy to be quite effective but most specify it is only effective if the patient has the mutation and some viewed the high cost of the medication as a disadvantage.
Our goal is to enhance the quality of lung cancer care through facilitating oncologists' practice according to not only evidence-based but also value-based individualized lung cancer care. The findings from this study will help us design effective implementation strategies for translating evidence- and value-based genomic medicine in clinical management of NSCLC in VA setting.
- Chen GJ, Arney J, Helm AM, Braun U, Richardson P, Chen P, Hayes TG. Genomic-based targeted therapy and management of advanced non-small cell lung cancer: Protocol for a qualitative study of oncologists' perceptions and behaviors regarding genomic-based targeted therapy. Journal of solid tumors. 2015 Apr 1; 5(1):doi: 10.5430/jst.v5n1p33.
- Arney J, Helm A, Chen GJ, Hayes TG. VA oncologists’ attitudes and behaviors regarding genomic-based targeted therapy for the management of advanced lung cancer. [Abstract]. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2016 Jun 3; 34(suppl):abstr e20573.
- Arney J, Chen GJ, Helm AM, Braun UK, Richardson PA, Chen P, Hayes TG. Oncologists perceptions and behaviors regarding genomic-based targeted therapy: Findings of a pilot study. [Abstract]. Journal of clinical oncology : official journal of the American Society of. 2015 May 13; 33(Supplement):e17560.